An Ergodic Theorem for Multidimensional Superadditive Processes

From MaRDI portal





The author studies ergodic theorems for processes \(F=\{F_ I\}\subset L_ 1(\Omega,{\mathfrak A},\mu)\) indexed by 2-dimensional intervals I in \(({\mathbb{R}}^+)^ 2\) or in \(\{0,1,...\}^ 2=V\). F is assumed strongly superadditive in the sense of \textit{R. T. Smythe} [Ann. Probab. 4, 772-782 (1976; Zbl 0339.60022)], with respect to a semigroup \({\mathcal T}=\{T_ u\}\) \((u\in ({\mathbb{R}}^+)^ 2\) or \(\in V)\) of linear operators in \(L_ 1\) with \(T_ uL^+_ 1\subset L^+_ 1\), \(\int T_ ufd\mu =\int fd\mu\) \(\forall f\in L_ 1\), and \(\| T_ uf\|_{\infty}\leq \| f\|_{\infty},\) \((f\in L_ 1\cap L_{\infty}).\) It is shown that \(u^{-2}F_{[0,(u,u)[}\) converges a.e. for \(u\to (\infty,\infty)\). The discrete case follows from a result of \textit{R. Émilion} and \textit{B. Hachem} [C. R. Acad. Sci., Paris, Sér. I 296, 85-87 (1983; Zbl 0537.47005)], who applied a reduction technique of \textit{A. Brunel} [Ann. Inst. Henri Poincaré, Nouv. Sér., Sect. B 9(1973), 327-343 (1974; Zbl 0272.47007)]. Instead, the author uses a technique of \textit{N. Dunford} and \textit{J. T. Schwartz} [Linear operators. I: General theory. (1958; Zbl 0084.104)] which is simpler in the author's opinion.











This page was built for publication: An Ergodic Theorem for Multidimensional Superadditive Processes

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3709524)